Excluded Volume of Polyelectrolyte in Salt Solutions

alent when eq. 8 is employed where Cc = Ch. How- ever, eq. 8 may have the added utility of providing a means of assessing the size of a “denaturing ...
0 downloads 0 Views 697KB Size
VOLUME OF POLYELECTROLYTE IN SALT SOLUTIONS

Feb. 20, 1964

alent when eq. 8 is employed where Cc = C h . However, eq. 8 may have the added utility of providing a means of assessing the size of a “denaturing unit,” given by the magnitude of N required to fit the data when cc # C h . Acknowledgments.-The authors wish to thank Dr. Milton Noelken for performing the optical rotatory dispersion measurements, Dr. Elkan Blout for graciously providing a sample of poly-L-glutamic acid, and Profes-

[CONTRIBUTION FROM

THE

543

sors E. Peter Geiduschek and Charles Tanford for enlightening discussions. We also wish to express our gratitude to Professor Stuart Rice for educating us on the proper choice of standard state for free electrons in aqueous solution. We gratefully acknowledge a discussion with Mr. Roland Hawkins, which gave us a better understanding of the content of ref. 18. The ideas in Appendix I resulted from the insight thus provided.

DEPARTMENTS OF APPLIEDCHEMISTRY A N D SYNTHETIC CHEMISTRY, NAGOYA UNIVERSITY, FURO-CHO, CHIKUSA-KU, NAGOYA, JAPAN]

Excluded Volume of Polyelectrolyte in Salt Solutions BY AKIRATAKAHASHI AND MITSURU NAGASAWA RECEIVED JULY 15, 1963 The intrinsic viscosities of sodium poly-(acrylate) having different molecular weights were determined in sodium bromide solutions of various concentrations. The analysis of the data was based on three different theories; Flory and Fox; Kurata, Stockmayer, and Roig; Stockmayer and Fixman. Both K. S.R . and S. F. plots converged a t one point, respectively, with different concentrations of sodium bromide, as those theories predicted, whereas F. F. plots did not give the same intercept against the prediction of the theory. From the slopes of K. S.R. and S. F. plots, the mutual excluded volume of segments was determined. The values obtained were compared with the values calculated using the Debye-Huckel theory for the potential of average force between two segments. Large disagreement was observed between the experimental and calculated results, and, moreover, the mutual excluded volume was found to be proportional to the inverse root of sodium bromide concentration, while the calculated result predicts its proportionality to the inverse of sodium bromide concentration. The reason for the discrepancy between theory and experiment is considered to be due to neglect of the counter-ion effect in calculating the mutual excluded volume.

Introduction During the past 15 years, great efforts have been devoted to clarify the well known expansion problem of a linear poly-ion in solution both theoretically and experimentally. The origin of the molecular expansion is now accepted to be the electrostatic repulsive force among fixed charges on the polyion. Owing to the electrostatic repulsion between fixed charged groups, the molecular dimension of the poly-ion increases to several times as large as that of the original uncharged polymer and varies markedly with the concentration of added neutral salt, ;.e., with ionic strength of solution. In spite of so many papers published to compute the problem quantitatively, there is still great disagreement between theories and experimental results. In all cases, the calculated expansions of the poly-ion are much larger than the experimental ones if we use the analytical charge density of poly-ion to calculate the repulsive force due to electric charges. The discrepancy is often explained by assuming that the effective charge density of the poly-ion is much smaller than the analytical one due to ion binding or low activity coefficient of the counter-ion in the polymer domain. To explain the discrepancy may be the most important problem to be solved in this field, but before going into the problem it seems necessary to re-examine the theory of hydrodynamic resistance used for polyelectrolytes, considering the recent progress in the theory on excluded volume of nonionic polymers. In most papers concerning the expansion of a polyion, the hydrodynamic resistance of the poly-ion is computed based on Flory’s theory.2 That is, the polymer coil is practically nondraining for solvent and the intrinsic viscosity is determined by the linear expansion of the polymer coil, a,,, as [q] =

KM1/~aV ( =S K M ” )

(1)

and

[?le

=

KM’/a

a7a = [93/[918

(3)

Here, [ q ] is ~ the intrinsic viscosity a t 0 temperature where the intrinsic viscosity is linear in the square root of molecular weight, M12, as shown in eq. 2. The intrinsic viscosity [ q ] is also related to the root-meansquare end-to-end distance (h2)‘/a by the equation3 (4)

where 9 is the well known universal constant of Flory. To relate this expansion facter a,, with molecular constants of polymer and solvent, Flory derived the equation

-

CY76

a?a =

cz

(5)

where C is a numerical constant and Z is a function defined by In eq. 6,a is the length of a link; M and ms,the molecular weights of polymer and segment, respectively ; and @ is the binary cluster integral defined by P =

-Jam

[l

- exp( - V(r)/kT)]4rr2dr

(7)

where V ( r )is the potential of average force between two segments as a function of their distance r . Although p is commonly used in this field, we prefer to use B converted from @ by the relationship B

=

P/2me2

(8)

for convenience of our discussion. Combination of eq. 5 and eq. 1 gives the following equation of Flory, Fox, and Schaefgena

(1) S. A. Rice and M. Nagasawa, “Polyelectrolyte Solutions.” Academic

Press, Inc., New York, N. Y.,1961. (2) P. J. Flory, “Principles of Polymer Chemistry,” Cornell University Press, Ithaca, N. Y.,1863, Chapter XIV.

(2)

That is

(3) P. J. Flory and T. G Fox, J. Am. Chcm. SOC.,78, 1904 (1861)

544

Vol. 86

AKIRATAKAHASHI AND MITSURU NAGASAWA

From the graphical plot of [ q ] z / 3 / J 1 vs. / S M / [v] one may evaluate both K and B , separately. Flory's theory was criticized by many a ~ t h o r sbe~.~ cause of the Gaussian distribution he assumed for the distribution of segments. After many discussions about the problem, Kurata, Stockmayer, and Roigj6 Stockmayer and Fixman,' as well a s p titsyn8 succeeded in deriving very similar equations of closed form, that is (Kurata, Stockmayer, and Roig6) 4/3g( ff )Z

cy3-,=

g ( a ) = 8a3/(3ffZ

+

(10)

1)1/%

(Stockmayer and Fixmanl) cy3

=

1

+ 22

(11)

where CY is the linegr expansion factor of the mean radius of gyration and is related with CY,, by4 ff,,3

=

ffz.43

If eq. 10 or 11 is combined with eq. 1, taking into account the ditkerence between CY and CY,,, we have (K. S. R.)

[?I

*/a/M'/a = Kz/a

+ 0.38300Bg(ff?)Mz/a/[?]'/a (12)

(S. F.)

[77]/dM =

K

+ O.blrPoBd/M

(13)

where CPO is the universal constant a t 0 point (-2.87 X 1021). As in eq. 9 we can obtain K and B from the graphical p b t s of [77]'/3/M'/avs. ~ ( c Y , , ) M [' /q~ p/ or [q]/da VS. 1 / M . Recently, Kurata and Stockmayerg reanalyzed the data of many polymers published and showed very clearly that eq. 12 gives much better agreement with experimental results than eq. 9. That is, the plots of experimental data according to eq. 12 give the same intercept a t Mf/8/[q]'/3 = 0 when various solvents are used for a polymer, whereas the plots according to eq. 9 do not give the same intercept a t M/[77] = 0 for different solvents. In,spite of the difference between models used in K. S. R. and S. F. theories, eq. 12 and 13 give almost identical results.' Equations 12 and 13 are very convenient for studying the expansion of polyions too, since we can easily obtain the mutual excluded volume of segments fl (or B ) from viscosity data, and also it is possible to discuss B from the standpoint of :electrostatic interaction between segments. There were some discussions published by Ptitsyn'O and Eisenberg and Woodside" along this line. Ptitsyn calculated B by assuming the DebyeHiickel ionic atmosphere around segments and inserted it into his equation, which is not shown here but analogous to eq. 10 and 11, for comparison with experimental results of carboxymethyl cellulose, poly(phosphate), and poly- (acrylic acid). Great disagreement was obserued, however, between calculated and experimental' results. Eisenberg and Woodside calculated the excluded volume function Z of poly- (vinyl sulfonate) from. the intrinsic viscosity data using eq. (4) M Kurata and H. Yamakawa, J . Chem. P h y s . , 18, 78.5 (1958); 19, 311 (1958); and also M. Kurata, H. Yamakawa, and H. Utiyama, Makromol. Chcm , 84, 139 (1959). (9) M. Fixman, J . Chem. F h y s . , 18, 1656 (1955); 86, 3123 (1962). (6) M. Kurata, W. H. Stockmayer, and A . Roig, ibid., 38, 151 (1960); and also M. Kurata and W. H. $tockmayer, Repf. Progr. Polymer Phys. J a p a n , 6 , 23.(1962). (7) W. H. Stockmayer and M. Fixman, J . Polymer Sei. Par1 C, No. 1 , 137 (1963). (8) 0. B. Ptitsyn and Y. Y . Eisner, J. Tech. P h y s . U.S.S.R., 119, 1117 (1959); and also 0 . B . Ptitsyn, VysokomolPkul. Soedin., 1 , 1200 (1959); S, 1673 (19fil). (9) IM. Kurata and W. H. Stockmayer, Forlschr: Hochpolymrren Forsch., 8 , 196 (1963). (IO) 0 . B. Ptitsyn, Vysokomolekul: Soedin., 3, 1084, 1251 (1961). (11) H . Eisenberg and D. Woodside. J. Chem. P h y s . , 8 6 , 1844 (1962).

10 and showed that Z/M'/' is independent of molecular weight of polymer. In this paper, eq. 9, 12, and 13 are compared by using the intrinsic viscosity data of sodium poly- (acrylate) of several differentmolecular weights in sodium bromide solutions of different concentrations. Much better agreement is observed in case of eq. 12 and 13 than in case of eq. 9. Moreover, from the slope of linear relationship of eq. 12 and 13 we evaluate B as a function of the concentration of sodium bromide and discuss i t based on the electrostatic interaction theory of electrolytes. Experimental Polymer Samples.-Sodium poly-(acrylate) used in this investigation was kindly provided from Toa Gosei Chemical Co. and its viscosity average molecular weight was 160,000. The sample was purified by precipitation from aqueous solution with methanol three ,times and was used for fractionation.I2 One hundred grams of the purified sample was dissolved in 6 1. of 0.4 N NaOH aqueous solution and the fractional precipitation was carried out at 25" through repeated stepwise addition of methanol-water mixture of 1 : 1 volume ratio which also contained 0.4 N KaOH. Twelve fractions were obtained, but seven fractions having suitable molecular weights were used for measurements. Each fraction was purified by precipitation from aqueous solution by adding methanol twice and washed with 90% methanol until excess NaOH was removed. The samples thus purified were washed with absolute methanol and ether, and then dried a t 50' in a vacuum oven for 2 weeks. When the samples were dissolved in pure water at about 0.1 N, the pH's of the solutions were 7-8. Molecular Weight Determination.-Molecular weights of these fractions were determined from the intrinsic viscosities a t 15" in 1.5 N NaBr solution ( 0 solvent13) using the equation [ 7 7 ] 1 . 5 = ~ ~ 12.4 ~ ~ ~X ~ 10-'MwOJ(15') (14) which was determined by using the light scattering method described previously. l 4 Molecular weights of each fraction thus determined are given in Table I.

TABLE I MOLECULAR WEIGHTSOF SAMPLES USEDFOR VISCOSITY MEASUREMENTS 7 -

FL4

FL3

FzL7

Fraction FzL6 F2L4

FiLl

FL1

M X lo-*

1 . 5 3 . 7 4 . 4 4 5 . 3 9 12.86 34.68 5 0 . 0 Reparation of Solutions.-Stock solutions of the fractionated samples and sodium bromide were prepared by weight using boiled distilled water; and the solutions having desired polymer and NaBr concentrations were made by mixing these stock solutions and boiled distilled water in volumetric flasks. Viscosity Measurements.-Intrinsic viscosities of the fractionated samples in NaBr solutions of various concentrations were determined with a capillary viscometer of modified Ubbelohde type, which had three bulbs for giving rate of shear corrections to the measured viscosities. The volumes of the three bulbs were; top 1.49; middle, 1.50; lowest, 1.47 ml. The heights of the centers of the bulbs, measured with a cathetometer, were 8.73, 5.81, and 2.93 cm., respectively. The capillary maintained horizontally had the length of 20 cm. and the radius of 0.039 cm., determined from the weight of mercury contained in a given length. The average rate of shear on a given solution, y , was calculated from the formula rhg k y=-=3Z~aVrel ?re1 with the assumption t h a t the densities of solution and solvent are the same. Here r is the radius of capillary, h the pressure head, g the acceleration of gravity, I the length of capillary, YO the kineI relative viscosity of solumatic viscosity of solvent and T ~ is~ the tion to t h a t of solvent. The factors k calibrated with the known kinematic viscosity of water at 15" were 469, 319, and 161 sec.-l for top, middle, and lowest bulbs, respectively. A few examples of the rate of shear dependence of sample FL1 in NaBr solution are shown in Fig. 1. The examples selected are the measurements where the molecular weight is the highest, the ionic strength is almost the lowest, and the concentration is also the highest. Therefore, the rate of shear correction was (12) A. Takahashi and I. Kagawa, J . Chem. Sac. J a p a n . I n d . Chem. Seclion, 64. 1637 (1961) (in Japanese). (13) A . Takahashi, S. Yamori, and I. Kagawa, ihid., 88, 11 (1962) (in Japanese). (14) A . Takahashi, T. Kamei, and I. Kagawa, i b i d . , 88, 14 (1962) (in Tapanese)

Feb. 20, 1964

VOLUMEOF POLYELECTROLYTE IN SALTSOLUTIONS

545

1

1 50 100 R A T E OF SHEAR (SEC:').

3~

I50

Fig. 1.-Relative viscosity a s a function of rate of shear for sample FL1 a t various ionic strengths of NaBr a t 15": 0 , C = 0, C = 0.0945 g./lOO ml., C." 0.126g./lOO ml., C." = 1 X = 5.02 X 0 , C = 0.0945 g./lOO ml., C." = 2.51 X

40 0.1 I

I

I

I

I

I I l l 1 1

I

I

1

I 1 Ill

IO0

IO

M x IO:* Fig. 3.-Intrinsic viscosity vs. molecular weights of sodium poly-(acrylate) in sodium bromide solutions a t 15'; ionic strength of NaBr, C.': 0 , 1.506; 0,5.02 X lo-'; 1.00 X 1 O - I ; A' 5.02 X V , 2.51 X a, 1.00 X 8,5.02 X u, 2.51 x 10-3.

30

+,

cj \

-E

&

F' 4 ,

od

20

0

1 cn t

IO

~

0

0.05 0.I C (Wioomt.)

Fig. 2.--Selected viscosity-concentration plots for sample FL1 a t various ionic strengths of NaBr a t 15": 1, C," = 2.51 X 2, C," = 5.02 X 10-3; 3, C," = 1 00 X 4, C," = 2.51 X unnecessary in most experiments other than such a few extreme cases (FL1 and F2L1, NaBr concentration 0.0025 and 0.005 N ) . The efflux times of water through these three bulbs a t 15' were 52.7, 80.3, and 154.0 sec., respectively. The kinetic energy correction factors were estimated from the efflux times of water a t 15 and 35", and found to be 0.67, 0.27, and O.S%, respectively. I n most experiments where the rate of shear correction was unnecessary, only the middle bulb was used, and hence the kinetic energy correction could be neglected. All measurements were carried out a t 15.0 f 0.01' for comparison with the data in '6 solvent.'3J4 AI1 solutions were filtered through a sintered glass filter of JIS No. 4 prior to viscosity measurements. Dilutions were made by volume in the viscometer. The flow times were determined to 0.1 sec. A few examples in determining [?] from vSp/C vs. C and In v r e l / C os. C under the most severe conditions are shown in Fig. 2. In experiments 1 and 2 of Fig. 2, the rate of shear correction was applied, while in others the correction was negligible. The values obtained for the intrinsic viscosity were found to be reproducible within =k5Y0 under such severe experimental conditions, but much better reprcducibility was found under ordinary experimental conditions.

Results Viscosity Equations at Various Ionic Strengths.All experimental results of intrinsic viscosities of the fractionated samples in NaBr solutions having different ionic strength are summarized in Table 11. In Fig. 3 are shown the graphs of intrinsic viscosities vs. molecu-

~ / c )x Fig. 4.-Flory,

IO:*

Fox, and Schaefgen's plots (eq. 9 ) ; ionic strengths of NaBr are the same as in Fig. 3.

lar weights in logarithmic scale. The lowermast curve in Fig. 3, which are the data a t the 8 point, is reproduced from previous paper^^^,'^ and the straight line has the theoretical slope of 0.5. As is clear from Fig. 3, the slope of log [VI V S . log M graphs increases with decreasing of ionic strength and seems to reach as high as 0.9, as expected from K. S.R. and S.F. theories. The values of K as well as v in the viscosity equation, [ q ] = KMY, calculated from Fig. 3 by the least square method, are summarized in Table 111. Considering the ambiguity in the experimental points a t low ionic strength, however, we should avoid concluding the limiting value of v a t low ionic strength. Comparison between Experimental Data and Theoretical Equations.-In Fig. 4, Flory, Fox, and Schaefgen's equation is compared with experimental results, and Kurata, Stockmayer, and Roig's plots and Stockmayer and Fixman's plots are shown in Fig. 5 and 6, respectively. Despite the prediction of these theories that all straight lines in each figure should give the same intercept, actually it is observed that straight lines in Fig. 5 or 6 give almost the same intercept-at least to the extent that straight lines drawn from the same

AKIRAT A K A H A S H I

546

AND

MITSURUN X G A S A W A

"0 -

Vol. 86

20

IO

0

X

L

-14

r/c;2

2

\

Fig. 7.-B vs. inverse root of ionic strength l/C,"l'z: 0, B obtained from S. F. plots; 0,B obtained from K. S. R . plots.

:I

h

v

from the slope of the straight lines in Fig. 5 and 6. The values of B determined from both theories are shown in Fig. 7 for comparison. Since the experimental points in Fig. 5 and 6 scatter considerably a t low ionic strength, it is true that the values of B obtained have considerable ambiguity a t low ionic strength, but a t high ionic strength the values are very reliable. The ambiguity in B thus obtained is indicated in Fig. ' i

Fig. 5.-Kurata, Stockmayer, and Roig's plots (eq. 12); ionic strengths of NaBr are the same as in Fig. 3.

TABLE I1 INTRIXSIC VISCOSITIESOF SODIUM POLY-( ACRYLATE) SOLCTIONS OF SODIUM BROMIDE AT 15" FL4

FL3

1.506 5.02 X 1.00 x 5.02 X 2.51 X 1.00 x

2

5.02 2.51

x x

10-1 10-1 10-2 10-2 10-2 10-3 10-3

0 . 1 4 9 0 238 ,222 385 358 655 438 84 1 550 I 029 689 1 68 1 . 0 8 2 29 1.51 2 74

DEPESDEXCE OF K moles/l.

1 .50414 5.02 X 10-l 1 . 0 0 x 10-1 5.02 X 2.51 x 1 0 P 1.00 x 10-2 5.02 x 10-3 2.51 x 10-3

I

nI

"0

I

I

2

I

I

I

I

4 M

I

6 ~

X

I

~

0 ~

Fig. 6.--Stockmayer and Fixman's plots (eq. 13); ionic strengths of NaBr are the same as in Fig. 3.

point do not look unreasonable-but the plots of Flory, et al., in Fig. 4 do not seem to converge a t the same point even though the ambiguity in experimental points is taken into account. This situation is the same as that observed for nonionic polymers, which was fully discussed by Kurata and S t ~ c k m a y e r . ~ Therefore, we may determine the value of K from the intercept and B

F2L6 F2L4 FiLl FL1 100 ml./g.-0.735 0.883 0.260 0.300 0 . 4 3 7 ,414 0 . 5 2 5 0.860 1.645 1 917 ,817 1.048 1 . 7 7 4.09 5.00 1.05 1 396 2 . 4 5 5.15 6.58 1.41 1 76 3.34 8.13 10 05 12.2 16.0 1 78 2.58 5 20 2.64 3 52 7.42 19.95 23.17 3 72 24.7 29.6 4.45 8.72 171,

TABLE I11 VISCOSJTYEQUATION ( [v] = KM") IOSICSTRENGTH AT 15"

AND Y I N ON

C.",

AQUEOUS

F2L7

-_

C,", moles/l.

IN

K X 10'

12.4 5.27 2.54 2.81 1.63 (1.36) (4.42) (2.49)

0.5 628 755

77 .84

.89 .83 .89

Here, it is to be noted that B obtained is proportional to the inverse root of ionic strength of solvents, l / d C z at least when the ionic strength is not too low, while Flory and OsterheldL5showed by using Flory's theoryI6 that B is proportional to the inverse of ionic strength, l/Cso. In our experiments, too, B would be proportional to the inverse of ionic strength if we evaluate B from Flory's theory at a constant molecular weight. Besides, i t is interesting to see that B a t infinite ionic strength is negative. The values of K obtained from Fig. 5 and 6 are almost equal and, hence, only the values determined from Fig. 6 are listed in Table IV. Strictly speaking, i t is observed that K depends slightly on the ionic strength of solvent. The unperturbed dimensions of polymers, (h02)1/s, can be calculated from K using the relationship (15) P. J. Flory and J . E. Osterheld, J . P h y s . C h e m . , 6 8 , 653 (lg54) (16) P. J . Flory, J . Chem. P h y s . , 21, 162 (1953).

VOLUMEOF POLYELECTROLYTE IN SALT SOLUTIONS

Feb. 20, 1964

547 (19)

The ratios of (hoz)>'/2to (bf2)'/z, the unperturbed dimension for the case of free rotation around individual chain bond, are also listed in Table IV and are found to be almost equal to those reported for nonionic polymer~.~ Intrinsic Viscosity of Polyelectrolyte a s a Function of Molecular Weight and Ionic Strength.-The inspection of Fig. 7 shows that B determined experimentally is expressed as B

=

Bo

+ Be = Bo + (A/&,')

(15)

where A is a numerical constant, Bo the value of B a t infinite ionic strength (that is, the term due to nonelectrostatic interaction), and Be is the term due to the electrostatic interaction. By inserting this equation into eq. 13, we obtain

[77]/dz=

K

+ 0.51(Po(Bo + A / d / C . " ) d / M

(16)

1 + (const.)z dF

TABLE IV UNPERTURBED END-TO-END DISTANCES AT VARIOUSIONIC STRENGTHS

1.506 5.02 X lo-' 1.00 x 10-1 5.02 X -5.02 x 10-3

K x 10' 1.24 1.5 1.6

1.6

[(ho')/ M 11'2 X 10"

[(hd))l(haf~)l"~

756 805 822

2.38 2.5 2.6

822

2.6

When Bo = -Be, we have B = 0, that is, the 8 point of the polyelectrolyte in salt solution. At this point, the osmotic second virial coeficient also becomes zero. l 3 Therefore, i t can be concluded that aqueous salt solution is a very poor solvent for the discharged poly-ion. Discussion The assumptions involved in the excluded volume theory of nonionic polymers are : (1) The free energy of one macromolecule in solvent is the sum of the configurational entropy of polymer skeleton and the interaction energies of all possible pairs of segments. ( 2 ) The potential of average force, V ( r ) ,rapidly drops off with distance between segments, and the mutual excluded volume /3 is significantly smaller than the volume occupied by a polymer coil, Le., /3